Please remove the column dupe. Existing files should be moved to names with the correct codes.
| langcode | dupe | # of files affected |
| lzh | zh-classical | 7 |
| nan | zh-min-nan | 1 |
| yue | zh-yue | 1 |
If you would like to discuss locally, please go to https://commons.wikimedia.org/wiki/Commons_talk:Timed_Text#duplicate_langcodes .
I only know of these three. Please check for others.
These three are dupes because they are identical to the correct codes. The rest of zh-xx are not dupes because they can be used to distinguish regional variants of Standard Chinese.
The builder for creating pages uses Xml::languageSelector and Language::fetchLanguageNames( $language, 'all' );
I don't think we have a way to filter out the dupes from there right now...
Regardless, you can simply rename the TimedText files as you see fit. It doesn't care about the particular code used, as long as it is valid.